Reliable protein-protein docking with AlphaFold, Rosetta, and replica exchange

Area of Science:

• Computational Biology
• Structural Biology
• Biophysics

Background:

• Predicting protein complex structures is crucial but challenging, particularly with conformational changes.
• AlphaFold-multimer (AFm) shows limitations in accurately predicting protein complexes, succeeding in only up to 43% of cases.
• Current methods struggle to model protein interfaces and predict complex structures when binding partners undergo significant conformational changes.

Purpose of the Study:

• To develop a robust computational pipeline for accurate protein complex structure prediction.
• To improve the sampling of conformational changes in protein-protein interactions.
• To enhance the prediction accuracy of AlphaFold-multimer for challenging protein complexes.

Main Methods:

• Combined AlphaFold (AF) for structural template generation with a physics-based replica exchange docking algorithm.
• Repurposed AF confidence measures (pLDDT) to estimate protein flexibility and docking accuracy.
• Integrated these metrics into the ReplicaDock 2.0 protocol to create the AlphaRED pipeline.
• Validated the pipeline on a curated collection of 254 protein targets and the Docking Benchmark Set 5.5.

Main Results:

• AlphaRED successfully docked 97 previously failed AlphaFold predictions.
• Achieved CAPRI acceptable-quality or better predictions for 63% of benchmark targets.
• Demonstrated a 43% success rate on challenging antigen-antibody targets, significantly outperforming AFm's 20% success rate.

Conclusions:

• Integrating deep learning (AlphaFold) with physics-based enhanced sampling (replica exchange docking) offers a powerful strategy for protein complex structure prediction.
• AlphaRED provides a robust in silico pipeline capable of accurately modeling protein interfaces and predicting complex structures, especially those involving conformational flexibility.
• The developed method successfully addresses limitations of existing tools like AlphaFold-multimer for specific challenging cases.